The purification of a liquid by reducing the concentration of ions or molecules in the liquid has been an area of substantial technological interest. Many techniques have been used to purify and isolate liquids or to obtain concentrated pools of specific ions or molecules from a liquid mixture. Known processes include electrodialysis, liquid chromotography, membrane filtration, reverse osmosis, ion exchange and electrodeionization. As used herein, the term electrodeionization refers to the process wherein an ion exchange material such as an ion exchange resin is positioned between anionic and cationic diaphragms. In contrast, the term electrodialysis relates to such a process which does not utilize ion exchange materials positioned between the anionic and cationic diaphragms. Although electrodeionization is quite effective in removing ions from liquid, it has never been developed to the degree that it adequately removes certain molecules or complexes that are large, heavily hydrated, weakly ionized or highly charged.
The first apparatus and method for treating liquids by electrodeionization was described by Kollsman in U.S. Pat. Nos. 2,689,826 and 2,815,320. The first of these patents describes an apparatus and process for the removal of ions within a liquid mixture in a depleting chamber through a series of anionic and cationic diaphragms into a second volume of liquid in a concentration chamber under the influence of an electrical potential which causes the preselected ions to travel in a predetermined direction. The volume of the liquid being treated is depleted of ions while the volume of the second liquid becomes enriched with the transferred ions and carries them in concentrated form. The second of these patents describes the use of macroporous beads formed of ion exchange resins as a filler material positioned between the anionic and cationic diaphragms. This ion exchange resin acts as a path for ion transfer and also serves as an increased conductivity bridge between the membranes for the movement of ions. These patents represent the primary structural framework and theory of electrodeionization as a technique.
Significantly improved electrodeionization systems have more recently been disclosed in U.S. Pat. No. 4,925,541 to Giuffrida et al. and U.S. Pat. No. 4,931,160 to Giuffrida, the teachings of both of which are incorporated herein in their entireties by reference. However, these systems still have the drawback of being unable to adequately remove certain molecules or complexes that are large, heavily hydrated, weakly ionized or highly charged.
The previous literature on technologies which rely on bipolar membranes and electroregeneration of resins does not discuss the role of membrane or resin crosslinking to any great extent. This probably is because it has been assumed that very high membrane permselectivities were required due to the large concentration gradient that occurs across the membranes during use. It also seems to have been assumed that, in the production of high purity water, (i.e. water that is both deionized and free of large or highly hydrated species), the transfer of ions was "film diffusion controlled" as in normal ion exchangers and therefore the diffusion rate of ions within the resin was unimportant. In film diffusion controlled systems, the controlling factor in transfer speed is the ability of ionic species to cross a "film boundary layer barrier" present at the interface between the resin and the liquid in which the ionic species resides.
The above assumptions have been strengthened by AC electrical resistance measurements of resins and membranes which show little difference in resistance in the presence of typical ions found in water. Thus, absolute resistance of components was considered of little importance because it related only to power consumption which is small for electrodeionization systems because they are used with high purity water applications having relatively low salinity as a starting point. Furthermore, electrodeionization was often used downstream of ion exchange softeners or scavengers which eliminates the need to remove large, highly hydrated or highly ionized species from feed streams using electrodeionization equipment. The drawbacks of electrodeionization systems resulted in the water treatment field developing relatively little interest in the removal of large, highly hydrated or highly charged species from feed water via electrodeionization.
The film diffusion model was strengthened in credibility when highly crosslinked resin beads having a substantially uniform diameter were substituted for conventional resins in electrodeionization apparatus. The use of such resins in electrodeionization apparatus is the subject of U.S. Pat. No. 5,154,809 to Orem, et al. This change resulted in improved performance due to an increase in the resin surface area and also due to an effective increase in the amount of resin active in the electrical circuit within the system.
The aforementioned problems with conventional electrodeionization apparatus can be exemplified using silica as a model. Compared with other dissolved materials commonly found in water, silica is present only in trace amounts. However, its removal becomes important in the production of high purity water; every trace constituent present in the feed water must be removed if the feed water is to be processed into a high purity form. It is well known that systems such as electrodialysis do not remove silica and that electrodeionization and electroregeneration techniques only partially remove silica. The inability to adequately remove silica from a feed water stream thus greatly reduces the applicability of the above techniques in high purity applications, including the largest high purity application, boiler feed water.
According to the "film diffusion" model discussed previously, silica is not well ionized and therefore does not transfer through the resin and membrane. However, such an assumption does not explain observations regarding performance of electrodeionization apparatus on silica removal. The change to resins having a substantially uniform bead diameter allowed complete removal of weakly ionized carbonic acid from the feed stream, but did little or nothing to help with removal of silica. However, when electrodeionization is operated at high voltage and/or low flow rates, silica can be picked up by the resin. As such, the silica must be in ionic form on the resin and must therefore have transferred successfully through the "film boundary layer barrier". Even in these cases, however, total transfer does not occur.
The literature has suggested pH adjustments of the water to a more basic form to ionize the silica and enhance its removal from feed water. However, such pH adjustments have been found to have only a moderate effect on silica removal. Instead of removing silica, the equipment rapidly removes the hydroxide ion that was added during pH adjustment and leaves the silica behind. Attempts to remove silica by bipolar electroregenration of resin have resulted in only partial removal as well.
Although exemplified by silica, many other trace constituents are not adequately or easily removed from feed water by electrodialysis, electrodeionization or bipolar/electroregeneration techniques. Some species diffuse so slowly or bind so tightly they use up available sites on the resin thereby reducing the capabilities of the resin to remove rapidly diffusing ions. The slow diffusing or binding species are referred to in the industry as "foulants". Although this problem is usually associated with anion resins, the problem is also known to occur in cation resins in the presence of multi valent cations.
Although also exemplified by silica, many constituents, particularly organic acids but also constituents like boron, have difficulty being removed from feed water because they are only weakly ionized. In this case, external means can be used to adjust the pH in order to provide higher ionization of the compounds. However, with existing devices using highly crosslinked resins, this proves to be an inefficient process because the ions to be removed still transfer more slowly than the added hydrogen and hydroxide ions needed to effect the ionization of the impurity. This condition results in excess electrical requirements or chemical feed rates.
In the case of silica, present electrodeionization devices operated under standard conditions are unable to remove more than about 80% silica under any conditions of voltage or current. As used herein, the term standard conditions for silica removal signifies a 13 inch flow path, nominal diluting and concentrating stream flow velocities, 5.degree.-35.degree. C., non polarity reversal, pH below 9 in the concentrating stream (either directly or via transport of hydroxide ions via the dilute stream), no addition of ionizing chemicals, no AC overlay, at least 1 ppm dissolved silica in the feed stream and steady state operation.
Sulfate removal using standard conditions has a limit of approximately 66%. For purposes herein, the term standard conditions for sulfate removal signifies a 26 inch flow path, single stage, 100 inch effective membrane area per cell pair, flow rate of 200 ml/(minute cell pair), 67-80% product water recovery, 15.degree.-25.degree. C., non polarity reversal, no AC overlay, steady state operation, approximately neutral pH and a feed stream containing a pure sodium sulfate solution of 250 ppm concentration. The equipment is operated at an applied voltage which provides electric current passage of about 1.25 amperes.
Under steady state conditions, present electrodeionization apparatus is also unable to remove highly charged (i.e. trivalent or greater) ions. For cell designs to date, voltage increases that effectively double the electric current across the cell have been sufficient only to increase the removal of highly charged ions by a factor of no more than about 25% of the incoming feed level.
Finally, present electrodeionization devices are unable to efficiently remove, under steady state conditions, large ions (i.e., ions having equivalent weights of greater than about 200). For present cell designs, voltage increases that effectively double the electric current across the cell have been sufficient only to increase the removal of such large ions by a factor of no more than about 25% of the incoming feed level.
The inability to completely remove silica, organics and certain metal ions and the increase in fouling sensitivity greatly reduces the applicability of electrodeionization and other processes in numerous water treatment applications. Therefore, as currently used, electrodeionization systems often require extensive pretreatment steps such as softening, organic scavenging, UV destruction of organics or reverse osmosis. They also often require post treatments such as ion exchange polishing. Additionally, in feeds containing mixed ions, electrodeionization cannot operate at maximum efficiency for high purity water production. These disadvantages increase the costs and complexity of a high purity treatment system thereby decreasing the competitiveness of the process over traditional techniques like ion exchange and reverse osmosis. The disadvantages also decrease the applicability of electrodeionization processes for water purification applications such as water softening, trace component polishing, electroplating, chemical synthesis and purification, food and beverage processing and waste treatment applications.